Text

STADIC4 Model for Frequency of Nonrecovery of Electric Power at Seabrook Station for Plant at Power & Shutdown
	ML20082E381
Person / Time
Site:	Seabrook
Issue date:	01/31/1990
From:	Kreslyon Fleming, Read J; PLG, INC. (FORMERLY PICKARD, LOWE & GARRICK, INC.)
To:	;
Shared Package
ML20082E364	List: ML20082E357; ML20082E381;
References
	PLG-0507, PLG-0507-R02, PLG-507, PLG-507-R2, NUDOCS 9107310383
	Download: ML20082E381 (30)
	v • d • e

-,

Rews'on 2 1

STADIC4 MODEL FOR FREQUENCY OF NONRECOVERY OF ELECTRIC POWER AT SEABROOK STATION FOR PLANT AT POWER AND SHUTDOWN l

w by Jack W. Read Karl N. Fleming Prepared for NEW HAMPSHIRE YANKEE Seabrook, New Hampshire January 1990 t

-lP G e APPLIED SC:ENTISTS e MANAGEMENT CONSULTANTS Neapon Bea-:n a CA ENGINEERS wasnegm a DC c'

__l t 1.

. p(

CONTENTS

1. INTRODUCTION 1

2. MODEL DESCRIPTION

'3

3. SEAGROOK RESULTS 18

4. COMPUTER MODEL FLOW DIAGRAM 23

5. COMPUTER LIST 26

6. REFERENCES 27 APPENDlX A. STATISTICAL DISTRIBUTIONS OF MODEL VARIABLES A1 APPENDIX B. PROB ABILITY DISTRIBUTIONS FOR ELECTRIC POWER RECOVERY SCENARIOS EPR 1 THROUGH EPR 40 B1 APPENDIX C. PICKARD, LOWE AND GARRICK, INC., lNTERNAL MEMO ON THERMO HYDRAULIC ANALYSES OF LOSS OF SHUTDOWN COOLING EVENTS C1 APPENDlX D. STADIC SUBROUTINE

SAMPLE

LIST D1 N PSNH1 N0022 012590 iii Pickard. Lowe.snd Garrict. Inc

T LIST OF TABLES 7

1.

Frequency, Durations, and Conditions of Outage 2.

Recovery f.iodels in Terms of the Availabihty of Emergency Feedwator and Recovery Possibilities for Seqnarios with the Plant in Operation 8

3.

Recovery Models in Terms of the Recovery Possibihties and RCS Condition for Scenarios with the Plant Shut Down 9

4.

Mean Probability c' Nonrecovery of Electric Power Comparison 19 5.

Reactor At Power 21 6.

Reactor Shutdown 22 LIST OF FIGURES 4

1.

Definition of Times for Events at Power 2.

Nonrecovery of Offsite Power Model 12 14 3.

Nonrecovery of Single Diesel Generator 4.

Nontecovery of Dual Diesel Generators 15 5.

Time of Core Damage after Onsite Power Failure with Plant in Operation 17 6.

Time of Core Damage after Onsite Power Failure Used for Seabrook PSA, Table 10.4 7 20 7.

Flow Diagram of Subroutine Sample 24 NPSNHI N0022 012590 iv Pickard, Lowe and Garnck. Inc

1. INTRODUCTION A calculational model was developed for the ST ADIC4 computer code to compute the conditional probabihty of failure of the onsite power system and no recovery of onsite or offsite electric power before the time of core damage. The model considers onsite power failures during the first 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months after loss of offsite power imtiating event. Although the frequency of loss of offsite power is excluded from this model, variation in its time of recovery is included. Recovery of onsite power is considered for any onsite power failures that occur during this 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months interval. The model was us?d to analyze scenarios with the plant in operation and during shutdown. It was also used to analyzed scenarios with the reactor having operated at 100% and 40% power. A large number of different plant stenarms is included to address the possibility that various initiating events may result in portions of the electric power system being nonrecoverable and to address different limiting conditions for the time available for recovery.

The model considers the impact of the depletion station electric batteries, two independent diesel generator trains, and the ability of the emergency feedwater pumps to delay the time of core damage. The time of core damage when the plant is n its operation configuration (i.e., modes 1,2, or 3) is taken to be the minimum time to severe core damage based on loss of coolant inventory timing from two competing leak paths: degraded reactor coolant pump seals when steam generator cooling is available and open power operated relief valves (PORV). The later path only applies to scenarios with no reactor coolant system heat removal and the plant in operation configuration. Severe core damage is taken as the time when peak fuel rod cladding temperatures reach 1,200*F.

For events initiated at shutdown in modes 4. 5, or 6, the time of core damage depends on the decay heat level, water inventory available in the reactor coolant system (RCS), and the RCS configuration. The configuration of the reactor vessel during shutdown (i.e., vessel head on or off) and the status of the residual heat removal system (RHR); i.e., RHR connected or isolated dictates the water inventory available for cooling and the leak paths available for coolant loss. The time of core unrecovery is calcuiated from the decay heat at the time o' station blackout and the water inventory available. The time of station blackout is taken as a I

random time within a variable duration outage.

Because of the uncertainties in the failure data and in the other parameters used in the model calculations, there is a resultant uncertainty in the frequency of nonrecovered station blackout. Uncertainty distributions are input into the STADIC4 code (Reference 1), and the electric power recovery factors nre computed in its subroutine SAMPLE for each random sample of parameter vahles selected from the uncertainty distributions. Hence, a probability distribution of the frequency of nonrecovered station blackout is calculated by the STADIC4 code for scenarios with the plant in operation or shutdown conflguration and having operat d at 100% or 40% power. Using this approach. the following sources of uncertainty are accounted for in the model.

Failure rates of electric power system components.

Maintenance frequencies and durations.

Common cause (beta) f actors.

Time to recover onsite power (diesel generators).

Time of steam generator dryout for different conditions.

NPSNH1N0022 012590 1

Pickard. Lowe and Garnck. In:

Time of core damage efter steam generator dryout.

Battery life after in tiation of blackout.

Frequency and duration of plant outages.

Fraction of outage time spent in different RCS configurations.

l l

NDSNH I Nn022 012f 90 2

Pickard. Lowe and Garrick. Inc.

2. MODEL DESCRIPTION The general equation for the probability (O) that onsite power is lost and that there is no recovery of onsite or offsite power before the time of core damage due to the station blackout during the first 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months after loss of offsito power is expressed as f(t)G f t)G (t )G (t;)dt (1)

Q=

i i j 2 t o where the probability of n ; offsite power recovery between the time Gj(t)

=

of offsite power loss (t ) and the time of onsite power failure (t).

o G j(t;)

the conditional probability of no offsite power recovery between the

=

time of onsite power failure at time (t) and the time of core damage (t + ;).

the conditional probability of no onsite power recovery between the G (T)

=

2 i time of onsite power failure at time (t) and the time of core damage (t + r;).

the probability density function for onsite power system failure; it is f(t)dt

=

the probability of onsite power failure in the interval (t to t + dt),

the time of onsite power failure.

t

=

Note that O is conditional on the occurrence of a loss of offsite power.

The time of core damage after onsite power failure at time t is the time tj (see Figure 1),

which is dependent on the emergency feedwater (EFW) pump wgrgigo; not working for the plant in operation scenarios. For simplicity, all EFW failures are4assumedito occur at t = t.o For events that occur at power, t is set at to = 0, and it is assumed that the plant operated o

continuously at a fixed power level until equilibrium decay heat levels were reached. For events at shutdown, t is taken as a random time between the time of shutdown from a o

previous equilibrium power condition and the end of an outage of variable duration.

For the scenarios with the plant shut down, the time, tj, is dependent on the decay heat level (at time after shutdown and the operation power level) and the water inventory to be boiled l

off. The water inventory available for cooling and the process for cooling is dependent on the shutdown configuration (vessel head on or off), the status of the RHR system (connected or isolated), and the steam generator shell side water inventory. The shell side water inventory of the steam generators provides additional heat capacity for those scenarios with the vessel head still on as natural circulation of the primary inventory will permit heat transfer to the steam p.enerator secondary, lo account for the possibility of maintenance, only two steam genera: ors are assumed available for this configuration. Then, both steam generator secondary inventory and the RCS Inventory are available to dissipate the decay heat by being boiled off through the primary PORVs and secondary steam dump valves. The path for primary inventory loss is dependent on the RHR configuration, if the RHR is connected and not isolated, the primary inventory would normally be lost initially through the RHR safety relief valves and then through the primary PORVs once the primary pressure exceeds the Nxwt uv22 ot2m 3

Pickard. Lowe and Garrick inc

-9 i

3 n

G z

c r__

e 5

d u

o*

3

%{

+3 IE

=

3

-g

s

-!l 0 -- LOSS OF OF FSITE POWER e

o

- LOSS OF ONSITE POWER m

<a3 m

~

BATTERIES RUN OUT F

m O

-4 5

A g

m 3-'

CORE DAMAGE u

I __ AN ALYSIS OF F ALLURES IS TRUNCATED x

2 E

ta O

T a

C) 2 s

a

RHR isolation setpoint (RHR isolates). However, for these scenarios analyzed, the RHR is always considered connected during shutdown because there is no power to isolate the RHR when station blackout occurs. RHR isolation normally occurs as a result of automatic closure of motor-operated valves (MOV) when RCS pressure exceeds 600 psia. From Appendix C.

the equation for the time to severe core damage following the loss of onsite power (without RHR isolation) and the vessel head on is given as:

_100% Power _Qases rj r 3.4[(T + Tsc)/24]0 " + Tsc i

40% power Cases tj = 7.86[(T + Tsc)/24]O " + Tsc where

= time after T when severe core damage occurs.

Tj i

= time after shutdown when loss of onsi*e power occurs.

Ti Tsc = time after T when steam generator dryout occurs i

and all times are in hours, in References 2 and 3, calculations are made to estimate T c as:

3 100% Power Cases Tsc = 2.38(T /24) 308 i

40% Power Cases Tsc = 16.08(T /24)'308 i

The time. T, is quantified based on the actual statistical data for outages of the Zion Nuclear i

Plant and is expressed as a function of the outage types (A, B, and C), outage conditions (W, X, or Y), the frequency of the outage type, the time required to get on RHR cooling, and the

' durations of the outage, which are dependent on the operation / maintenance procedures require.d for the outage type.

The type of outages considered are: (A) nondrained maintenance outages, (B) drained maintenance outages, and (C) refueling outages. The three types of outage conditions that can apply are: (W) RCS is closed, water inventory is full, and steam generator cooling is available: (X) RCC is open and the water level is between the vessel flange and midplane of steam generator nozzles; and (Y) RCS is open and water level is at the refueling level. Only outage condition W can apply for type A outages. Both W and X conditions can apply to type 8 outages and all three conditions (W, X, Y) can occur during type C outages.

The duration for each type of outage is determined by the number of operation / maintenance procedures that must be followed for each type of outage and the statistical duration time determined for being in each procedure in addition to the time after shutdown to get on RHR cooling. Table 1 indicates the variables used in the model to calculate the time of each type of outage it also shows the variables for the procedure times involved and the outage condition for each procedure. The frequencies determined for each type of outage and the 64PSWiN0022 01:599 5

Packard Lowe and Garrick Ine

duration for being in each procedure tree are input to the model as uncertainty distributions.

These distributions are based on the maximum, minimum, and average values determined for the Zion Nuclear Plant, as derived from Reference 4.

The model was developed to analyze the various scenarios shown in Tables 2 and 3, which are the possible combinations of available recovery functions and plant conditions. The different recovery possibilities are dictated by whether offsite power is recoverable, one or two dieset generators are recoverable, and the emergency feedwater system is working during the station blackout event. These factors are explicitly modeled elsewhere in the Seabrook risk model (Reference 5), Therefore, the offsite power recovery model is evaluated separately for the 40 different casos identified in Tables 1 and 2. For a given accident sequence, only one of these models is applied.

By defining the following terms-l F = the frequency (fallures per demand) that one specific diesel generator fails to start 3

and the other succeeds when called on at time t = to (taken from the Seabrook systems analysis).

F2 = the frequency (failures per demand) that both diesel generators fail to start when called on at time t = to (taken from the Seabrook systems analysis).

l = the diesel generator failure rate during its first hour of operation.

i 12 = the diesel generator failure rate during its 2nd through 24th hour of operation.

= the conditional frequency that the cause of failure to operate for one diesel generator will be shared by the other diesel generator (the 8 factor for failure to star 1 is implicitly accounted for in the terms F3 and F )-

2 Tne probability density function for onsite power failure, f(t)dt, can be developed to account for the possibilities of failure to start if two diesel generators are available for recovery from f(t)dt = [(1 - 2F1 - F )(22e ' - (2 - )leq2-St) + 2F le"' + F ] dt (2) 2 i

2

.iy eo where 2 = l or 2, depending on the time interval.

i 2

In the above expansion of f(t), the first term accounts for the case when both diesel generators start successfully, for which the frequency is (1 - 2F - F h the second term i

2 accounts for the case when one diesel starts and the other fails, for which the frequency is 2F ; and the third accounts for the case when both fail to start, for which the frequency is F -

2 j

The term (2)e# - (2 - B)le42-Wt) is the probability density function for onsite power failure when both diesels start. The term le* is the density for the failure of the single diesel that started successfully. The density when both fail to start is 1. The reason for the above expansion is to facilitate the integration of Equation (1) in light of the discontinuities at t = t.

o Substituting Equation (2) into Equation (1), the conditional probability (Q) of nonrecovery of onsite or offsite electric power in 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months is expressed as N#5NH1Nn022 n125M 6

Pickard. Lowe and Garrick. Inc

Table 1. Frequency, Durations, and Conditions of Outage Duration (outage conditions)

Outage Outage Time To Procedure Trees Type Frequency go on RHR 1

2 3

4 5

6 T A6(W)

A FA TAl(W)

T A1(W)

TBS (W)

TB6(W)

B FB T Bl(W)

TB1(W)

TB2(X)

C FC l

TCl(W)

TCi(W)

TC2(W)

TC3(Y)

TC4(X)

TCS(W)

TC6(W)

NPSNHIN0022 012590 7

Pickard. Lowe and Garrick. Inc

g awiiis--

isis lable 2. Recovery Models in Terms of the Availability of Emergency Feedwater and Recovery Possibilities for Scenarlos with the Plant in Operation E nmgene F Electric

Power Recovery Function Available" b

Recovery Model One of Two One Offsite Yes No Diesels Diesel Power EPR1 EPR 2 EPR-3 E PR-4 EPR5 E PR-6 EPR 7 EPR 8 EPR-9 EPR 10 EPR 1 through 10 models for reactor having operated at 100% power; EPR 31 through 40 models are the same as shown except for the reactor having operated at 40% power.

" Recovery function availability: * = available; blank = unavailable.

)

Npenu t Non?? ni? con 8

Pirb aed Lowe and Garr6 Imr

Table 3. Recovery Models in Terms of the Recovery Possibilities and RCS Condition for Scenarios with the Plant Shut Down Electric

Power Recovery Function Avaltabie" RCS Recovery One of Tw I5It' Conditio. t One Diesel 0

Diesels Power W

EPR-11 W

EPR-12 W

EPR-13 W

EPR 14 W

EPR 15 X

EPR 16 X

EPR-17 X

EPR 18 X

EPR 19 X

EPR 20

EPR 11 through 20 shutdown models for reactor having operated continuously at 100% power: EPR 21 through 30 models are the same as shown except for the reactor having operated at 40','. power.

" Recovery function availability; * = available; blank = unavailable.

tW = RCS closed with two steam generators available for cooling; X = RCS open with water level between need flange and hot leg midplanes.

l l

l NPSNHI N0022 012N 9

Pickard. Lowe and Garrick, inc

O = F G (T )C (T )

2(1 - F - F )

2 1 1 2 j i

2

! = to t = to

~

to 41 10 +24 d

t e G (t)G (tj)G (T )dt + );

e 2 G,(t)Gj(tj)G (!,)dt (3)

=x 1

i 3

2 s 2

3

,

to

10+I

~

't e"G (t)G (r;)G (t )dt + J o +24 "2'G (t)G (tj)G (T )dt

'to +1

--(2-J'1 - 2F - F )

2 i

i 2 j 2

e i

i 2 j i

2 1

,

to

to +1 where Tj depends on whether or not the emergency feedwater pump is working and time t.

For the scenarios with the plant in the shutdown configuration, tj depends on the time af*er shutdown when the loss of of' site power occurs, T, and time t.

i The equa%n tot the conditional probability (Q) of loss of onsite power within 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months of loss of offsite power and nonrecovery of onsite or offsite electric power before core damage, if one diesel has failed as a result of the initiating event and only one diesel generator is available for recovery, is O = F Gj(t))G (T ) + (1 - F )

i 2 J i

t=0t=0

~

to+1 to+24 -;2 0 (t)G (rj)G (Tj)dt t

i'G (t)G (rj)G (Tj)dt + ).2 J

e e

3 3

2

=

3 i

2 i

,

to

to 41 where the tj dependency is the same as stated above.

If offsite pcwer is unavailable for recovery, then the terms G (t) and G (r) are equal to 1.0 i

i (i.e., nonrecovery is guaranteed) and Equations (3) and (4) can be simpilfied.

If both diesels have failed and no dit.,el generators are available for recovery, then the equation is expressed simply as I

(5) 1 Q=

where r;is determined at time t =0 and depends on whether or not the emergency feedwater l

pumps are working or the time after shutdown. The above cases enable the application of the model to scenarios in wb 1 some external everit or hazard results in a nonrecoverable loss of offsite or onsite power.

The computer code STADIC4 was used to calculate the uncertainty distributions for Q, the frequency of onsite power failure and distribution of Q, the frequency of onsite power failure and nonrecovery of onsite or offsite electric power before the time of core damage, given the initial loss of offsite power at t = to and the loss of onsite power at any time in the interval (0 to 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months ),

ajocilu t ajnnoo ni? con 10 Pickard. Lowe and Garrick Inc

equations described above were programmed into the ST ADIC4 subroutine S AMPLE The a'Ong with centrol parameters to describe the models for nontecovery of offsite power, nonrecovery of a single diesel generator and nonrecovery of a dual diesel generator, shown in Figures 2. 3. and 4, respectively. The uncertamties in these models are shown on each figure in terms of the 5th,50th, and 95th percentiles of an uncertainty distribution.

The curves II' Figure 2 were developed from Public Service of New Hampshire (PSNH) data on duration of outages on individual transmission hoes. The 5th percer, tile curve i:. based on the assumption that the recovery from all three independent transmission lines coming into seabrook is the same as for an individual line. The conservative 95'h percentile is developed assuming that the PSNH line data can be applied independently to all three lines-The 50th percentile curve is the average of the 5th and 95th percentile curves The curve in Figure 3 is based on an analysis of generic data on diesel generator recovery, which indicates that about ore-third of these failures are nonrecoverable and that, of those that are recovered, most of these are recoverable within about 6 hours6.944444e-5 days 0.00167 hours 9.920635e-6 weeks 2.283e-6 months . The curves in Figure 4 were developed from Figure 3, using a procedure similar to that described above for Figure 2.

Since the recovery of onsite and offsite electric power is dependent on the availability of DC power, the uncertainty of the lifetime of the plant batteries was modeled as having a probability distribution of 0.05,0.80, and 0.15 for a battery life of 2.0,5.5, and 9.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months , respectively (Reference 5). The recovery of offsite power without casite power (diesel 1

generators) available was assumed to be delayed 2.0 hours0 days 0 hours 0 weeks 0 months , once the batteries have failed, to allow time to make a decision about which breakers to close, the control room supervisor briefing of auxiliary operators, the auxiliary operators manually closing the breakers, and correcting breaker malfunctions (or choosing alternate paths or a set of breakers). No uncertainty was defined for this conservative assumption of 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months in this analysis.

However, the program could be easily modified to account for this uncertainty by replacing this constant by a value with an appropriate uncertainty distribution.

The 100% power cases EPR-1 through EPR 20 that were originally reported in Revision 1 of this report have been rerun with the updated distributions for offsite power recovery (Reference 6) and for the life of the plant batteries (Reference 7), The new results are shown in Taoles 5 and 6. The updated distribution for the nonrecovery of offsite power is shown in Figure 2 (Revised), which shows the uncertainty distributions in terms of the 10th,50th, and 9Cth percentiles. A four-point histogram is used to represent the updated battery life uncertainty, which is based on calculations of limiting battery life performed by Yankee Atomic. The new discrete uncertainty distribution used,has battery life of 2.9, 4.47, 4.60 and 4.71 hours8.217593e-4 days 0.0197 hours 1.173942e-4 weeks 2.70155e-5 months , with the probabilities of 0.00045,0.14955, Oj0, and 0.15, respectively. The 40% power cases EPR-21 through EPR 40 were not rerun since the 40% power limit operation of Seabrook is currently not being pursued. If interest develops again for the 40%

power limit operations, these cases should be terun with the new offsite power recovery and battery hfe distributions.

Since DC power is also required to recover the diesel genera' ors, no credit was taken for diesel generator recovery after battery depletion.

The time of core damage (t) after onsite power fails is dependent on the status of the EFW pumps when tha plant is in operation. If emergency feedwater is available, the time of core damage will be extended because the core after heat continues to be removed and core l

NmW196622 n12",0 11 hckard. Lowe and Garrick Inc

l

N

_ N

- O N

~

- @ d "C

D 9

w w.c

~

3d<

w "C

"a

~

w 0

5o I

O u

w

- e-m e

w o

s=

c.

w w

u*

e z

w 2

I w

- cc -

o u

w C.

a 5e w

.t v

- o o

S cw O

D

~

O

- v

>ou O

w C

O N

z d

G

,s I

I f

y g

c.

e.

~

e.

m.

n.

N.

P u.

ADN3nO3Ud 83 mod 30 A83 AOD38NON 3 A11V70WnD s

I

J et *Ju t *Jnn?? n1254n 12 Peckard. Lowe and Garrick, Inc

l1

,lj

ll M

1' #

i 22 0

i 2 s

i i s

i i

, E I

l 1

lit l

lA m

F H

E W

2 O

I i

P l

l lE T

l F

e A

d E

o o

M i

I M

l I

re wo t

^

P E

L

{a e

1 1

t r

i E

s f

C f

E R

O t

t E

iJ i

o t

P f

l l

t n

I4 i

o y

f t

g 6 E

C r

l l

e C

lE v

l

'P o

i i

i t

c t

tt 0

e 5

r

/

no I 4 N

)des iv l 2 e

R

(

2 e

ru

,I o

g iF 3

2 s

4 0

9 a

7 6'

i hNw$ aW bI0 Y<y

,a/3 w2 5

l 3

2sa E: $a ?f; _

l l

l jl

I t

1.0 5

.g

?

A o

.8 z

wao

.7 u.

Ew3 i

o

.G Sill. 50Til. 95Til PE RCE N TIL ES a

u_o

}

y

.5 m

O>

0 Ow

.4 a

h Ez O

n z

.3

(

w 2e<

.7 2

3o

.1

=

f.

I I

I I,p 0

1 2

3 4

5 6

24 g

t 11ME AFTER DIESEL FAILURE.f TOURS

$a o

l 5

5 Figure 3. Nonrecovery of Single Diesel Generator n

^ ^ ^ - '*M

F LJ Al E

B to C

u wd c

4 w

y 4

O A

w U

c D

f e

e u

N

{

D

=

0 E

I 2

~i w

e a2 O

M J

g O

uo 5

t2 E

O v

O 5

~

E b

W f

E O

o t

t O

N E

E i

i s

i e

I e

e.

~

x.

n.

~

3 ADN3nD383 W3 mod 30 AW3 AOD3BNON 3Al1Y10 WOO

[

NosNw t Non22 012590 15 Pickard Lowe and Garrick. Inc.

coolant inventory is only lost via the degraded reactor coolant pump seals, which fail at time

/

g, Withou EFW pumps working, the coolant inventory is lost via the coolant pump seal leakages, which start at t = 0, as well as through the primary coolant relief valves (PORVs) because there is no core decay heat removal from the reactor coolad systern, The models for the time of core damage following tro loss of onsite power, with and without emergency feedwater pumps available, are shown in Figure 5 for scenarios with the reactor having

{

operated at 100% power. The uncertainties of the models are shown as t,,e 5th,50th, and 95th percentiles, which were assigned probabilities in ST ADIC4 of 0.2,0 6, and 0.2, respectively.

When the reactor has operated continuous at only 40% power, the decay heat is sufficiently reduced such that the time of core unrecovery without EFW pumps working is dictated primarily by the coolant inventory loss via the pump seal leakages. Hence, for the 47,.

power scenarios with the plant in operation, the curves of Figure 5 for EFW pumps working are applicable to both scenarios with and withou* EFW pumps working, The uncertainty of nonrecovery models in Figures 2,3, and 4 were modeled in STADIC4 with probability distributions of 0.1,0.8, and 0.1 for the 5th 50th, and 95th percentile uncertainty curves, respectively.

The uncertainly and statistical distributions of all of the variables used in the computation model are provided as STADIC4 input and shown in Appendix A.

Using the Monte Carlo simulation technique, the STADIC4 computer code is used to select a single random number for each variable based on the probability distributions provided as input, For each set of selected varltbles, the probability distribution of the nonrecovery of electric power is calculated by STADIC4 for each power recovery scenario (EPR 1 through EPR-40), The calculations are repeated as many times as necessary by STADIC4 to obtain a statistically acceptable probability distribution, l

I l

l NPsNWI N0022 012590 16 Pickard Lowe and Garrick inc

32, A

30 E

2e

/

26 7

/

24

,e g

5 e'

O

/

G

/

p 5

22 c

4 Y

/

~

/[

~

p.3

/

,g

+

/

30 Ot 18 pp l

r 16 8E E

l l

14 wu

/

s i

,2.-

kb w

10

/

p O

I

/

i l

/

g t

g

/

EVERGENCY FEE 0 WATER PUMPS WORKING E

j 6

/

- --- EMERGENCY FEEDWATER PUMPS NOT WORKING

/

AT TIME 1

/

4

/

l

/

I 1

l 2

s I

I I

?

0 O

2 4

6 6

10 12 14 16 16 20 22 24 TiVE WHEN ONSITE POWER 15 LOST. HOURS Figure 5. Time of Core Damage after Onsite Power Failure with Plant in Operation NDSNW1 Nnn22 012590 17 Pickard Lowe and Carnck inc f

3. SEABROOK RESULTS A

The Origin I analysis provided in the Snabrook probabilistic safety assessment (PS A) for the conditional frequenc y of power f ailure at bus ES or "

on a loss of offsite power initiating event (the noreec overy of electric power), was shc ction 10 4 of the Soabrook PS A E

peference 5). The mean vah;es for scenarios EPR.

Jh EPR-10 frorn that analysis, are shown in Table 4 a'Jng with the mean values calcul ath the c omputer mndol describod in this repcut, but with sonie of tho niajor roodol m sunytions of the original analysts, whit h verifies the calculational methodology of this computer model. The results calculated with the models and assurnptions of this report are ako shown in Table 4 for comparison The i

I rnaior chfferonce between the original analysis :,nd the analysis with the current rnodol of thm report is in the rnodel used to calculate the tirne of core darnage after loss of onsite electric power. The original analysis considered no uncertainties in the model and was based on an earlier core heatup analysis. Figure 6 shows the time of core damage (r) model originally used, which can be comi ared to the current model shown in Figure 5 that is based on more current core heatup analyses.

The mean values of the results calc.ilated with the models described in this report for 100% power,40% power,100% powtr shutdown, and 40% power shutdown (scenarios EPR 1 through EPR-40) are provided in Tables 5 and 6. The results for the 100, power cases have been calculatet.' using the updated distributions for offsite power recovery and expected plant battery life, av discussed in Section 2.0. The 40% power shutdown results are those from Revision 1 of this report, calculated with the earlier distributions of offsite power recovery and Oattery hfe; they would need to be rerun if there is resumed interest in the 40% power cases.

The probabihty distributions for the electric power recovery scenarios. EPR-1 through EPR-40, using the current model described in this report and the STADIC4 computer code, are provided in Appendix B.

1 t

e, 1

1 r

l

. t 4

1 l

l l

4 NPSNNI N0022 01:090 18 Peard Lowe and Garre inc

=

I 4

Table 4. Mean Probability of Nonrecovery of Electric Power Comparison ea r A

STADIC4 Calculan Scenario wtA Poe Table 10.4-7 with Table 10.4 7 76.lels

[- l Uncertainties.A y s

EPR 1 264 184 2.2 4 E PR-2 5.1 4 8.1 4 7.7-4 EPR 3 2.0 3 1.9-3 2.4-3 EPR 4 5.6 3 8.0 3 7.6 3

\\

EPR 5 4.1-3 4.9-3 4.5 3 EPR 6 4.6 3 7.2 3 6.2 3 EPR 7 3.7 2 3.7-2 3.6 2

[

EPR 8 4.2 2 5.2-2 4.9-2 EPR-9 7.5 2 8.2 2 1.3 1 EPR 10 2.9-1 2.9 1 2.8 1 NOTE: Exponential notation is indicated in abbreviated form; i.e.,2.6 2.6 x 10" l

l 1

I N PSNH1N0022 012$90 19 Pick ard lowe and Carrick inc

~.

=

24 j

22

~

/

20 1

18

/

D0 16 c:

14 o

U=

D w

12

=

8 0

1 10 1

w G

/

j B

EMERGENCY FEEDWATER PUMPS WORKING 8

EMERGENCY ' EEDWATER PUMPS NOT WORKING

/

4 2

I l

0 0

2 4

6 8

to 12 14 16 18 20 22 24 TIME WHEN ONSITE POWER IS LOST, HOURS Figure 6. Time of Core Damage after Onsite Power Failure Used for Seabrook PSA, Table 10.4-7 i

NPS NW1 Nn022 012590 20 Pdard Lowe and Gerick 'a*

Table 5. Reactor At Power Electric Number of Recoverable Offsite Power Emergency FW Power Diesels Recoverable Available

^"

Recovery p,

Model 0

1 2

Yes No Yes No (New) 100% Power Cases i

1 EPR-1 8.60-5 EPR-2 4.92-4 EPR-3 9.27-4 E PR-4 4.99 3 EPR5 5.33-3 E PR-6 6.65-3 EPR-7 4.08 2 EPR-8 5.10-2 EPR-9 4.95-2 EPR 10 1.93-1 1

40% Power Cases l

EPR 31 2.21 E-4 EPR 32 2.21 E-4 EPR-33 2.39E-3 i

EPR 34 2.39E-3 i

EPR-35 4.52E-3 EPR 36 4.52E-3 EPR 37 3.61 E-2 EPR-38 3.61 E-2 EPR-39 1.30E-1 EPR-40 1.30E 1 NOTE: Exponential notation is indicated in abbreviated form; i.e., 8.60-5 = 8.60 x 10 5 l

l l

N PS NH1 N0022 012590 21 Pick a d Lowe and c.vei-k w

Mle 6. Reactor Shutdown 7lectric Number of Recoverable Offsite Power Power Diesels Recoverable RC Mhs*

P a

h Recovery Model 0

1 2

Yes No (New}

1000.'c Power Cases EPR-11 W

4.53-2 EPR-12 W

3.78-3 EPR-13 W

2.67-4 EPR 14 W

1.01-1 EPR-15 V/

1.47-2 EPR-16 X

8.20-2 EPR-17 X

1.34-3 EPR 18 X

9.75-5 EPR-19 X

3.90-2 EPR 20 X

4.69 3 40% Power Shutdown Cases EPR-21 W

5.42E-2 EPR-22 W

4.23E-3 EPR-23 W

3.37E-4 EPR-24 VV 9.22E-2 EPR-25 W

1.28E-2 EPR-26 X

6.39E-2 EPR-27 X

1.14 E-3 EPR-28 X

8.81 E-5 EPR-29 X

3.50E-2 EPR-30 X

4.00E-3

W = RCS closed with two steam generators available for cooling; X = RCS open with water level between head flange and hot leg midplanes.

NOTE: Exponential notation is indicated in abbreviated form; i.e., 4.53-2 = ~.53 x 10 2 lP 1

I

4. COMPUTER MODEL FLOW DIAGRAM

/

The computer model for rionrecovery of electric power developed for the STADIC4 computer code is written in fortran computer languar and provided as subroutme SAMPLE. A flow diagram of this subroutine is shown in Figure 7.

f

\\

l i

4 i

i l

I 1

l 1

t i

I

'e NPSNHIN0022 012590 23 Pickard. Lowe and Garrick. Inc

gwet a hauPJ h

9 c -

.,8

,',Te ri,t,..,. 3a

-4=-

~ '~ -c e,ae,. ea par 4*sa.Fv =0f s e

,g3 t

wwaste ta nt av

?a' a s ' a ' t ut * ' t v,,t N,ggfry La sem 08' t,'t a=0 cg D4Sf't*O*t#

eess g etc.3,gav cces, t s 9

la**Fonsa=0 m

,,g g,,

, _ 7,

-.0 St* men 6 wawtS I

9 ww *,' t c 0ee,W* a,'c,g t' w g a

SET D'm*af Qvi n, 1"4 8

rvo o, c,

80*E Last %t ait e vts yg,,,g,

g 8ett*B4/'T seOts N #vW

? W asec'us C. s t't

'en e.-

9

O

<p g

68T 0 e acCovtar, coessvg ca mt t sectatuTV seDU CowaVt atices ccnapVTatton nese t Ors e a f3 YestS t'O YOt 900Ci T C 'Ot T000 80m 88tC4aDJPV seetta 8CamitDh tVPTceTs g

- e <*wo ow1merissue swresoPwpsom SET h4MaY LF8 I

PN WeDO t e s ta

,gg g

TO&avt 90

?

SE1 taseagg.sc, M Nw g

88tD&tLifT PC1N suw. a se g

1 N

1powta e

'f8 oP*t"ccytn.

ma serv. g to, tog MT DC.av f est POR 4ed 08'co'vgssy apega ll 180wta g

. =0 i

et I

tafftesta Paa,

Of LAv. 3 e saa m

g

,gg st?

g

-'0'.

10. s

0. s. m..

..,,g,,,,,,

a.e.,on a.$

r 4F'I A tasVMDewes k iso h uC Muse m a t snare

DespV'1 GtfC CO**N'EC*8'O Lvwevtsge*o g

" E * 'o * 'O.

soa t o,, vo.

O pon TO e tt aaso DELAf ani gg j,

,9

,g mEC3vtPPV GOS #as

%,,,ta m "IW "' IOS ' B5 atOOvt n' ECS ici ao, g,, %

'EEe W aseO t 9

9 Y

'0D%. *wt T O er% = %e at? f rist' O

't3 GSTOs e 08'e

70. 70tav E

3EY h

'.81T. aspe. g M*O 00usNt a teose gg esOCES 90a6

Eedto SCEhamen 4 soo -

Sff SET #804assufT OF 0*et D 4 paeL3 toe 700C

'O ST aarf 8ee ais 9

1Q sti panneaec

wat es gr>'m D G twL i

to g'agit

91. s *e S Medt

'f5 9

9 w' o e a es saa.

wi o g au= sat.

v t 88 shaft a=0 st*a vat shatt amo tr's eacica som ga mA

+

e ac*om soa ?3mm I6.ne,ana a ra M, a.a.. a n a

.,a.

a o

6 Figure 7 (Part 1 of 2). Flow Diagram of Subroutine Sample NPSNHIN0022 012590 24 Pickard Lowe and Garrick Inc

.m

.. = _. _.., _ _

'O t

i s

D setas h

4us et m

.I 83 #

tune eset h3 I. OF 4 D G.e6 0 e

esc rEet e et

$9 ggy h

W ll to, Me tWPW te.pa.a e. 2 j

It D gov. bat t

. /:...., N.,

N

, /

...eesaeD3 N.

.. /

r,,.

-I

's e

mc.

4,,.

.c

..s

..s N E0'9 S.fn.eP.UT. acT.

p

-=

=

g;ea,

=..

+

.e -.

~~

~ ~,,.,

-.e..

e e m...

.n.e essaT S.C sa up e

.e...c,.

e.e.

9

, - =,%

..e 4..

'.L*'_-

< '.. - ~ =

v

..t.

=.e 8'CM

.D.o.t.e N.T 9

.~

Figure 7 (Part 2 of 2). Flow Diagram of Subroutine Sample NPSNHI N0022 012590 -

25 Pickard. Lowe and Garrick Inc.

.,.. _. _ _ _ _. _... - _...,.. _.,,,., _, _ _. _.,, _ _ _...... ~.. _

I

5. COMPUTER LIST i

The computer list of the fortran program subrcutine for the nonrecovery of electric power model iised with the STAD!C4 computer code is provided in Appendix D.

1 N PSNH1 N0022.012590 26 Pickard Lowe and Garrick. Inc

l 6.

REFERENCES i

1, Pickard, Lowe and Garrick, Inc., "STADIC4 Computer Cnde User Manual," PLG-0438, September 1985.

2.

Pickard, Lowe and Garrick, Inc., " Thermal-Hydraulic Analyses of Postulated Loss of Decay Heat Remo.>el Events during Shutdown," prepared for Public Service cf New Hampshire, PLG-0595, December 1937.

i 3.

Pickard, Lowe and Garrick, Inc., " Thermal-Hydraulic Analyses of Postulated Loss of Decay Heat Rernoval Events during Shutdown Following Long-Term Operation et 40% Power," prepared for Public Service of New Hampshire, PLG-0601, January 1988.

4.

Pickard, Lowe and Garrick, Inc., and Commonwealth Edison Company, " Zion Nuclear Plant Residual Heat Removal PR A " prepared for Electric Power Research Institute, NS AC-84, July 1985.

5.

Pickard, Lowe and Garrick, Inc., "Seabre ok Station Probabilistic Safety Assessment,"

PLG-0300, prepared for Public Service Company of New Hampshire and the Yankee Atomic Electric Company, December '983.

6.

Pickard, Lowe and Garrick, Inc.,"Cata for the Frequency and Duration of Loss of Offsite Power Events at Seabrook St: bon," prepared for New Hampshire Yankee, PLG-0726, December 1989.

7.

"An Evaluation of Seabrook Station Battery Life After the Loss of AC Power," letter from J. Stetkar, Pickard, Lowe and Garrick, Inc., to L. Rau, dated December 27,1989.

I NPSNH 1N0022 012590 27 Pickard, Lowe and Garrick, Inc.

.

ML20082E381

Text

Navigation menu

Search